2016 INFORMS Annual Meeting Program

WC67

INFORMS Nashville – 2016

WC65 Mockingbird 1- Omni Health IT Sponsored: Information Systems Sponsored Session

2 - Optimize The Prognostic Performance Of Data-level Fusion Approach Changyue Song, University of Wisconsin - Madison, Madison, WI, United States, csong39@wisc.edu, Kaibo Liu Multiple sensors have been widely employed to monitor the degradation status of a unit. Since each sensor signal contains partial and dependent information, data- level fusion methods have been proposed to combine the information and make decision on the remaining useful life. Existing data-level fusion methods are mainly based on experience and still lack theoretical basis to ensure their prognostic performance. To fill the literature gap, this paper proposes a novel data-level fusion approach with the following features: 1) It ensures the prognostic performance; 2) It allows varying parameters for different units; 3) The training and testing procedures are integrated. 3 - Optimal Reliability And Commonality In Component Design: A Service Logistics Perspective Geert-Jan van Houtum, Eindhoven University of Technology, Eindhoven, Netherlands, G.J.v.Houtum@tue.nl, Joni Driessen, Joachim Jacob Arts An Original Equipment Manufacturer sells multiple types of machines to groups of customers under full service contracts. For one component, we consider the design choice between one common component for all machine types and a dedicated component per machine type. We compare the Life Cycle Costs and find general conditions for the optimality of choosing a common and dedicated component, respectively. For a third area, numerical work has to show what is optimal. We also show how these areas behave as a function of numbers of sold machines and the commonality cost factor. 4 - Economic Design Of Run To Failure Binary State K Out Of N Systems With Identical Components And Geometrically Distributed Lifetimes Tugce Martagan, Eindhoven University of Technology, t.g.martagan@tue.nl, Murat Kurt K-out-of-N systems represent integrated collection of N individual units of which at least K of them are required to function for the system to survive; and have broader ability to model the operating structure of complex industrial systems. We consider the design of run-to-failure K-out-of-N systems consisting of identical components with geometrically distributed lifetimes under total and average cost criteria. We propose exact and faster approximate approaches to minimize the total cost due to operation, design and failure of the system and derive managerial insights from numerical analyses. WC67 Mockingbird 3- Omni Decision Models in Risk, Reliability and Maintenance Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Adiel T. de Almeida-Filho, Assistant Professor, Universidade Federal de Pernambuco, Caixa Postal 7471, Cordeiro, Recife, 50630971, Brazil, adieltaf@googlemail.com 1 - Data Fusion For Degradation Modeling In Multi-edged Machining Processes Wenmeng Tian, Virginia Tech, 250 Durham Hall, Blacksburg, VA, 24061, United States, tian0414@vt.edu, Jaime Camelio Tool wear is a widely used direct degradation measure of machining process, and can be measured from the cutting edges using machine vision system. Though there are methods available for single-edge machining process, degradation modeling for a multi-edged machining process using mixed-type data is still a challenge. This work proposes a novel scheme for degradation modeling and prognostics of multi-edged machining process by fusing mixed-type sensor data, including images and process signals. A broaching process is used as a case study to illustrate the effectiveness of the proposed approach. 2 - An MCDM Model For Maintenance Outsourcing Felipe Macedo de Morais Pinto, Universidade Federal de Pernambuco, Recife, 50630971, Brazil, felipe_mmp94@hotmail.com, Adiel Teixeira De Almeida Filho This work presents an MCDM approach for a maintenance outsourcing decisions, which including contract selection and supplier selection. Given the MCDM nature of maintenance outsourcing, models to address this problem include maintainability, dependability, quality of repair besides cost, which are detailed in the reference Multicriteria and Multiobjective Models for Risk, Reliability and Maintenance Decision Analysis.

Chair: Changmi Jung, Johns Hopkins University, Johns Hopkins University, Baltimore, MD, 21202, United States, changmi@jhu.edu 1 - Analyzing Interactions Between Social And Technical Systems:

A Retrospective Study In The Pediatric Environment Yi-Chin Kato-Lin, Hofstra University, Hempstead, NY, United States, YiChin.Lin@hofstra.edu

Using the lens of Socio-technical theory, we examine the effect of digitization of individual pain plans (IPP) on process outcomes of care delivery for patients with sickle cell disease, and the interactions between digitized IPP and care delivery teams. We found improved care coordination and satisfaction, unaltered usage behavior among physicians, and an unexpected, negative opinion among nurses that might explain the neutral/negative process outcomes after the digitization. Our results provide new insights into the important and inconclusive debate about the diverse impacts of HIT. 2 - Dynamics Of Social Influence On New Employees’ Use Of Volitional Is: m-EHR Case In Hospital Setting Sanghee Lim, The Johns Hopkins University Carey Business School, 100 International Drive, Baltimore, MD, United States, lim.sanghee@jhu.edu, Juntae Kim, Byungtae Lee, Jae-ho Lee, Yura Lee We investigate the effect of social influences on m-EHR usage by new doctors who recently began working at a hospital. Drawing upon the concept of organizational socialization and social influences, we hypothesize that coworkers’ m-EHR usage is positively associated with one by new doctors, and the strength of this association varies by the coworkers’ type of usage, by the hierarchical rankings of coworkers, and by the stage of socialization process in which the new doctors are situated. Our analyses using longitudinal m-EHR usage data generally support our hypotheses. 3 - Does Health Information Technology Change The Prospect Of Continuity Of Care? Changmi Jung, Johns Hopkins University, changmi@jhu.edu Continuity of care has long been the core element in healthcare discussion particularly concerning the improvement of care quality and patient health outcome. However, the emergence of healthcare information technology (HIT) is changing the landscape. As an enabler to comprehensive data storage and immediate information retrieval, HIT is filling the gap in the degree of available information, and thus, empowers more seamless care from end to end. In this study, we examine the effect of HIT on the patient health outcome and the extent of continuity of care in the hospitals in Maryland. Also, we explore the impact of the change in the degree of continuity of care on the patient health outcome.

WC66 Mockingbird 2- Omni Novel Approaches in Reliability Modeling and Analysis Sponsored: Quality, Statistics and Reliability Sponsored Session

Chair: Tugce Martagan, Eindhoven University of Technology, P.O. Box 513 5600 MB, Eindhoven, N/A, Netherlands, t.g.martagan@tue.nl Co-Chair: Murat Kurt, Merck & Co., Inc., TBD, North Wales, PA, 19454, United States, murat.kurt7@gmail.com 1 - Reliability Sensitivity Analysis Based On Kriging Metamodel And Importance Sampling Juan Wang, Nanjing University of Science and Technology, dongzao1115@gmail.com Due to the lack of the gradient information, the reliability sensitivity analysis can’t be implemented in an analytical manner.Therefore we proposed a simulation- based sensitivity method. It is suggested to compute the failure probability by the combination of Kriging model and importance sampling at first, and then the estimator of failure probability is differentiated through the score function approach. The calculation of failure probability resort to the construction of an accurate Kriging surrogate of the limit state function, and score function approach enables the estimation of the gradient of the failure probability without any additional evaluation to the limit state function.

447